Inspiration
As avid learners teachers and universities cannot always guide the way, we find ourselves in the age of infinite resources we still can't help but feel the resource curse. We decided to solve this problem using Iskul.
What it does
Iskul curates a learning path for each learner based on their expertise, time they can invest.
How we built it
We use SOTA LLM models like LLAMA and Whisper for inference using Groq hardware accelerators. However, we don't leave everything to AI, we fetch tutorials made by people.
Challenges we ran into
First challenge was the unavailability of video transcripts. Secondly, longer transcripts led to higher latency due to increased network calls. Variable lengths of transcripts need to be handled at the time of LLM usage.
Accomplishments that we're proud of
- Crediting the original content creators for their work.
- Orchestrating multiple AI tools to work in tandem.
- Dynamic UI for AI apps. ## What we learned
- Collaboration with AI tools .
- Team work.
- Faster development and deployment. ## What's next for Iskul
- Better UI/UX framework.
- Better accuracy of the content generated.
Log in or sign up for Devpost to join the conversation.